Camelia, the Perl 6 bug

IRC log for #bioclipse, 2010-01-18

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
07:44 egonw joined #bioclipse
08:08 Gpox joined #bioclipse
08:11 egonw Gpox: ping
08:12 egonw Gpox: someone asked about a GridGenerator, and asked about how to know what lines to add...
08:12 egonw do you have ideas on that?
08:12 egonw most important bit is how to practically get the visible screen coordinates, and how to convert those back to world coordinates...
08:13 egonw so that the Grid IGenerator can know what lines to put in that world space...
08:13 egonw but perhaps you have other ideas... ?
08:25 Gpox AtomConatinerBoundsGenerator
08:27 Gpox we are talking of multiple molecules, IGenerator could be made generic and you put in a AtomcontainerSet instead in a AtomContainerSetRenderer
08:28 Gpox why is the visible coordinates important we don't do any such culling of other generator elements
08:30 egonw because the grid drawns lines (rectangular) and one needs to know where they are drawn...
08:30 egonw on the IGenerator level... are there alternatives then?
08:36 Gpox one can pass in a object at the construction of the generator to hold the information
09:01 egonw could you send a reply with your ideas to cdk-jchempaint@ ?
09:06 mgerlich joined #bioclipse
09:15 olass joined #bioclipse
09:15 olass moin
09:15 zarah hello olass, you fantastic person you
09:15 zarah olass: You have new messages. Write '/msg zarah @messages' to read them.
09:16 egonw hi olass
09:16 jonalv joined #bioclipse
09:17 olass hi egonw
09:17 olass @msg
09:17 zarah olass: 3 d 21 h 53 m 32 s ago, jonalv asked should the splash screen in master say 2.3.0?
09:17 * olass is back from a nice holiday with skiing
09:17 jonalv sounds nice
09:18 olass very nice indeed
09:19 * olass is reading all emails now
09:19 olass takes quite some time to browse only a few days
09:39 sneumann__ joined #bioclipse
13:02 masak joined #bioclipse
13:11 samuell joined #bioclipse
14:03 egonw joined #bioclipse
14:42 egonw olass: I noted the bloody long list of 2.2 things on my name...
14:42 egonw Mylyn was quite capable in hiding most of them :) mylyn--
14:42 egonw but going through the list right now...
14:42 olass egonw: yes indeed
14:42 olass sounds good
14:42 egonw thanx for the 2.2 update
14:43 olass sure
14:43 olass most is just to move
14:43 olass to 2.4
14:43 * masak never used Mylyn, but just got another justification not to
14:43 olass but good to review things
14:43 * olass likes mylyn
14:43 masak I like the web interface.
14:44 masak if there was a pure-text terminal interface, I'd probably use that :)
14:56 jonalv masak: if you configure it wrong then it sucks. If you configure it right it is quite nice
14:56 masak that may be so. I've just never seen the initial need to do bug tracking from within Eclipse.
14:58 jonalv masak: well you want to do everything from within Eclipse of course
14:58 jonalv hm or at least some people want to
14:58 jonalv I just helped Jarl with Bioclipse
14:58 masak I just know that sometimes I actually prefer to go to vim or the shell to do things.
14:59 jonalv It seems sdf is a very well used file format
14:59 jonalv I think maybe we need a dialog or wizard for creating an sd file from pasteing
15:00 jonalv that is a dialog where to paste the sdf file content
15:00 jonalv as it is now you have to:
15:00 jonalv 1. create an empty file and name it .sdf
15:00 jonalv 2. close the empty molecules editor that opened
15:00 jonalv 3. open the file with the text editor
15:00 jonalv 4. paste
15:00 jonalv 5. close the editor
15:00 jonalv 6. open the file usieng the molecules editor
15:01 jonalv and then you are ready to use it
15:01 jonalv does anyone know an easier way of doing this?
15:03 jonalv anyone? egonw? masak? olass?
15:04 olass no, that sounds like a logical workflow
15:04 olass why is that wrong?
15:05 jonalv olass: I will suggest another
15:05 olass if you have text to paste, this procedure seems suitable
15:05 Gpox pasting in the navigator like eclipse, it depends how good the format recognition is
15:05 olass Gpox: aah, right
15:05 olass select the navigator and just paste creates a new file with the content
15:05 jonalv oh cool
15:05 olass that should be the way to go
15:05 jonalv with what name?
15:06 olass unknown.sdf?
15:06 olass unnamed.sdf?
15:06 * jonalv gives this a try
15:06 olass NewSDFile.sdf
15:06 olass jonalv: I have a feeling you might need to implement something.. but it should be doable
15:07 olass good luck
15:07 jonalv hm nothing happens
15:08 jonalv okey tried in Eclipse
15:08 jonalv seems they are doign some fancy hacks
15:09 jonalv maybe we should have a paste dialog instead?
15:09 egonw you know there is a New SD File wizard, right?
15:09 jonalv it's a question of discoverability
15:09 jonalv egonw: oh hang on does it do exactly this?
15:11 jonalv egonw: no it does not, this was a use case when I had the sdf file content on the clipboard
15:11 egonw no, it's not exactly that use case...
15:11 egonw but that one is easy, not?
15:12 egonw just create a new File, open it in a text editor, and c/p
15:12 jonalv egonw: uhm well yea I ran Jarl through it for a couple of files and it is a 6 step process...
15:12 egonw ah, ic
15:13 egonw but, why would you need to copy/paste anyway?
15:13 egonw can't you just save in the other tool?
15:13 jonalv I would have prefered !. open dialog 2. paste. 3 done!
15:13 jonalv egonw: yea we probably could but the other tool was running inside Windows and finding those files from Bioclipse would probably not have been easier...
15:14 jonalv Bioclipse was in OS X
15:14 egonw ah, ic...
15:14 egonw yes, lovely OSs they are :)
15:14 jonalv (and yea parallels++ does clipboard between OS X and Windows (so cool))
15:14 jonalv it does not do such fancy stuff for Ubuntu though :(
15:15 egonw I was refereing to their ideas of file systems :)
15:15 egonw and interoperability
15:15 jonalv ah yea that...
15:15 jonalv egonw: it\s a part of the system and not wroth fighting against ;)
15:15 egonw jonalv++ (for reading recommendations ;)
15:16 jonalv anyways I was thinking that a more general solution to this kind of problem might be a filr from paste dialog
15:16 jonalv I like the Eclipse solution but I don't think it is the Bioclipse way
15:17 jonalv to low discoverability
15:17 egonw yes, c/p normally works on the active editor...
15:17 egonw but there is not really a short cut for just opening a text editor...
15:17 egonw or a general c/p handler that would decide what editor to open it in...
15:17 jonalv maybe we could create a new dialog that recognizes what is on the clipboard and creates a suitable file?
15:18 egonw or that
15:18 jonalv how intuitive is it to just paste in the Navigator?
15:21 jonalv egonw: do we have code for guessing file formats?
15:21 egonw the Nav would need focus
15:21 egonw yes
15:21 egonw exposed in teh CDK manager
15:22 jonalv I vote for 'New -> File from clipboard' and 'Edit->Paste as file' both doing the same thing
15:23 egonw sounds like an interesting feature...
15:23 jonalv egonw: so identifying chemistry files is not difficult at all?
15:24 jonalv masak: what is your feeling regarding recognizing bioinformatics files?
15:24 masak jonalv: could you be more specific?
15:25 masak in general, I'm for it.
15:25 jonalv masak: I mean how hard do you think it is? What help can we get from , say Biocjava?
15:25 masak oh, Biojava does it a great deal already.
15:25 masak I delegate all file reading/recognition to biojava in the aligner.
15:26 masak s/aligner/sequence editor/
15:26 jonalv masak: so it's rather easy in fact?
15:26 masak aye, but hard to tell what the limits are.
15:42 edrin joined #bioclipse
15:46 jonalv egonw: I am guessing I should cdk.fromString. I am trying to track down what it does if it does not recognize the file format now...
15:46 jonalv egonw: do you know that behavior by hear perhaps?
15:47 egonw jonalv: have a look at cdk.determine*Format()
15:47 egonw various flavors
15:48 jonalv egonw: oh I was looking deep in CDK code. Seems it returns 'null' if it fails to find anything. (I should have guessed that)
15:49 jonalv egonw:  so in other words we wil lget a NPE if we fail to identify the type?
15:49 * jonalv tries to confirm that hypothesis
15:50 egonw could be
15:52 jonalv yea it returns null but we don't get a NPE but a BioclispeException complaining about not being able to create a reader for null
15:53 jonalv When methods return null anything can happend later on...
15:53 egonw yes, yes... write a unit test and file a bug report
15:54 jonalv egonw: will do
15:56 egonw > pubchem.search("HIV")
15:56 egonw List<Integer> of size 15
15:56 egonw ?
15:56 egonw and how can I see what the content of the List is?
16:02 jonalv egonw: for( Integer i : list ) {js.say(i);}?
16:02 egonw yeah, lovely...
16:03 egonw I rather just see the content as returned value... :(
16:03 jonalv well you will have to take that up with olass I am pretty sure it was his feature request, oh and I think masak did the job :)
16:03 * jonalv is not sure what he preers
16:03 jonalv |prferes
16:03 jonalv |prefers (gah)
16:03 * jonalv blames everyone today
16:04 egonw :)
16:04 masak correct -- I made that change.
16:04 masak olas has already retracted it a bit.
16:04 masak we're refining it on a case-by-case basis.
16:04 olass I think we need a way to annotate that we want to present all results
16:04 olass yup
16:04 olass indeed
16:04 * jonalv has to leave soon to pick up his new drivers license
16:04 masak I don't particularly like the annotation idea.
16:05 masak I think we're missing something.
16:05 * olass is happy with whatever solution
16:05 masak perhaps we should have a character limit on the returned values.
16:05 * jonalv would liek to hear everyones needs and then try to suggest a solution
16:05 olass yes, perhaops that is best
16:05 masak jonalv:  fair enough. I don't have any needs in this case.
16:05 jonalv exaclty olass: Tell me your problems! :)
16:05 egonw and perhaps a js.say(List) method
16:06 masak oh please no.
16:06 * olass does not have a working Eclipse anymore :(
16:06 egonw masak: now I', definately missing something here
16:06 egonw I'm
16:06 masak print and say are already polymorphic on their parameters.
16:06 * jonalv does not have a working MySQL anymore meaning he can't work with Brunn
16:06 masak why add more variants rather than fix the underlying problem?
16:07 jonalv masak: could you explain what the underlying problem is?
16:08 masak well, the original problem was that people who used the js console a lot, like olass, got too much output from some calls.
16:08 masak a list of sequences, for example.
16:08 masak by the way, no js.say is involved in this case.
16:08 olass sequences can be very long
16:08 olass and very many
16:08 olass and ouputting 100 sequences of 20K each is not desirable
16:08 olass floods the console
16:08 masak we just happen to stringify sequence objects to their actual sequence, regardless of length.
16:09 masak maybe that is the underlying problem.
16:09 jonalv masak: that is probblay not a good idea
16:09 masak on the other hand, sometimes that comes in handy.
16:09 egonw I use the console very much and get too little output now
16:09 masak but it feels like an easy fix to limit the output of the sequence objects.
16:09 masak nod
16:09 * egonw mubles config options..
16:10 egonw max var lengths before truncation
16:10 egonw max list lengths before 'list of length foo'
16:10 olass I think I installed a Java update for Mac and now I cannot run anything in Eclipse :(
16:10 egonw olass: what java version did it install?
16:10 olass no idea
16:10 masak I'd much prefer good defaults in this case to config options.
16:10 CIA-37 bioclipse.cheminformatics: Egon Willighagen 2.2.x * r14eeffe / plugins/net.bioclipse.pubchem.tests​/src/net/bioclipse/pubchem/tests/Ab​stractPubChemManagerPluginTest.java : Added missing unit tests for the download methods - http://bit.ly/7ECAZs
16:10 CIA-37 bioclipse.cheminformatics: Egon Willighagen 2.2.x * ree238e8 / plugins/net.bioclipse.pubchem/src/net/bioc​lipse/pubchem/business/PubChemManager.java : Fixed name of the implementation, to match the interface - http://bit.ly/7Xaw2d
16:10 * olass suspects this is the error anyway
16:11 jonalv masak: yea good defaults would be _so_ much better
16:11 egonw masak: you never use good default values for options??
16:11 masak they're a first priority, at least.
16:11 masak I'm not saying I don't want conifg options.
16:11 masak I just don't want them as much as better defaults.
16:12 egonw and we all know how much everyone agrees on default max lengths...
16:12 masak generally, I have a slight disincentive towards config options.
16:12 egonw like for source code lines, e.g. :)
16:12 egonw 'good value' is a good word
16:12 egonw or 'good default'
16:13 masak egonw: I could easily code in a project where the preferred max length of source code lines were something other than 80. just give me a value, and I'll go with it.
16:13 * egonw is slowly becoming allergic to the use of the word 'good' in science
16:13 masak smaller than 20 would be difficult, though :)
16:13 egonw there is no 'good' or 'bad' in science
16:14 egonw 'good' is meaningless
16:14 masak that's what I like about it.
16:14 egonw sure, it relieves you from any responsibility
16:14 masak I can be a mad scientist and do evil things, and no-one will stop me.
16:14 egonw as the famous Dutch (current) prime minister
16:14 masak he's a mad scientist?
16:14 egonw "if we knew then what we know now, we would have done it otherwise"
16:15 egonw (in reply to approving UK/US to invade Irak, without any legal and/or factual basis)
16:15 jonalv egonw: oh I am sure the word 'good' has some meaning just not a meaning which is so crystal clear and well defined as we sometimes would hope it to be
16:16 egonw well, the problem is more that it is not defined at all
16:17 egonw it's a label, an end product, not a target
16:17 egonw let's make Bioclipse good
16:18 jonalv egonw: yea but how? (do we make Bioclispe good?)
16:18 masak by making it better!
16:18 egonw no clue, that's the point indeed :)
16:18 masak (duh)
16:19 egonw better actually is better...
16:19 egonw still not useful in itself...
16:19 egonw but at least it suggests that there is some way of *comparing* what is good and what it not
16:19 egonw well... what is good and what is better :)
16:22 masak if 'good' isn't good but 'better' is better, what about 'best'?
16:22 masak should we make Bioclipse best?
16:22 masak it does sound attractive...
16:22 * jonalv grabs masak and heads off to pick up a new drving license and then go swimming
16:22 masak I... hey!
16:29 egonw grmph... bloody #pubchem is spitting bad MDL molfiles...
16:29 egonw lovely
16:30 egonw hack upon hack upon hack
16:30 egonw welcome to the age of semantic computing ...
16:35 CIA-37 bioclipse.cheminformatics: Egon Willighagen 2.2.x * r20abcb6 / plugins/net.bioclipse.pubchem/src/net/bioc​lipse/pubchem/business/PubChemManager.java : Fixed downloading as 3D: call correct 3D asString method (fixes #1803) - http://bit.ly/7WKDQE
16:35 zarah bug #1803 | http://tinyurl.com/ycfay43
16:35 CIA-37 bioclipse.cheminformatics: Egon Willighagen 2.2.x * rb093c6e / plugins/net.bioclipse.pubchem/src/net/bioc​lipse/pubchem/business/PubChemManager.java : Added workaround for PubChem spitting MDL SD files, not MDL molfile, and moreover, with erroneous training empty lines - http://bit.ly/603oHC
16:45 CIA-37 bioclipse.cheminformatics: Egon Willighagen 2.2.x * r2f475e3 / plugins/net.bioclipse.cdk.business/src/ne​t/bioclipse/cdk/business/CDKManager.java : Throw an exception when the input format is not recognized (implements #1813) - http://bit.ly/7pP6Zp
17:09 sneumann__ joined #bioclipse
17:47 edrin left #bioclipse
19:25 sneumann joined #bioclipse
20:13 ilogger2_ joined #bioclipse
20:15 ilogger2 joined #bioclipse

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary